Hyaluronan and its cell surface receptors, CD44 and RHAMM, are involved in the cascade of events that lead to growth and metastasis of several tumor types. Transfection of tumor cells with soluble, hyaluronan-binding forms of CD44 or treatment with hyaluronan oligomers inhibits murine mammary carcinoma or melanoma growth in vivo, as well as formation of metastatic nodules in the lung after dissemination from the vasculature. Thus hyaluronan interactions are an essential contributor to progression of at least some tumor types and introduction of reagents that perturb hyaluronan interactions is likely to be therapeutically effective. Thus the major objective of this proposal is to design recombinant adenoviruses driving efficient production of hyaluronan-binding polypeptides that effectively inhibit tumor progression, but are engineered to have minimal side effects by removing non-hyaluronan-binding domains to whatever extent is possible without losing efficacy. The relative efficacy of recombinant adenoviral constructs encoding hyaluronan-binding polypeptides composed only of hyaluronan-binding domains (i.e., link modules), or of hyaluronan-binding motifs within a neutral background, will be compared to constructs encoding soluble CD44's that contain more extensive extracellular domains, in the murine mammary carcinoma system used previously. Successful constructs will then be tested in vivo for their ability to inhibit rat gliomal invasion and to allay human breast and ovarian carcinomas, grown in nude mice, at different stages of progression. Such constructs could be used widely for experimental manipulations in numerous systems and may ultimately lead to useful strategies for gene therapy in cancer patients.